JP6623932B2 - Delivery schedule creation program, delivery schedule creation method, and delivery schedule creation device - Google Patents

Description

  The present invention relates to a distribution schedule creation program, a distribution schedule creation method, and a distribution schedule creation device.

  In a wide area distributed environment, large amounts of data may be distributed to a plurality of nodes. Hereinafter, a device that provides data to be distributed is referred to as a file server (FS), and a data distribution destination is referred to as an edge server (ES). For example, tens of GB or more of large-capacity data may be distributed to tens of thousands of ESs installed at a plurality of bases.

  When large-volume data is distributed to a plurality of ESs, distribution scheduling is performed in advance. In the distribution scheduling, the order of data distribution is determined so that the time until data transmission for all distribution requests received from data users is completed (total transmission time) is as short as possible.

  In order to efficiently perform data distribution, multi-stage distribution using an ES that has received data as a distribution source has been considered. In the multi-stage distribution, data is transmitted from the FS to the ES as the first-stage distribution. In the second and subsequent stages, the ES that has received the data becomes the distribution source, and transmits the same data to the nearest other ES requesting the same data.

  As a technique related to the efficiency of data distribution, for example, there is a multicast data distribution method that reduces the frequency of retransmission processing in multicast distribution and improves the speed of the entire distribution processing. Further, when distributing a large amount of data to a large number of terminals using a communication line, a data distribution device capable of completing distribution to all terminals in a short time has been considered.

JP 2000-286845 A JP 2000-22686 A

  In the case of performing multi-stage distribution, the time until the distribution is completed is affected by the communication speed (band) of the communication path between the distribution source and the distribution destination. In the search for the distribution schedule that can complete the distribution in the shortest time, the total distribution time is calculated for all patterns in the distribution order for all combinations of the distribution source and the distribution destination. Such a calculation corresponds to a combination optimization problem. In a large-scale network having a large number of servers, a combination explosion occurs, and it is difficult to perform the calculation within a practical time.

  In one aspect, an object of the present invention is to reduce the creation time of a distribution schedule that enables efficient data distribution.

In one proposal, a distribution schedule creation program that causes a computer to execute the following processing is provided.
The computer includes a data providing device that holds data to be distributed, a plurality of destination devices to which the data is distributed, and a plurality of relay devices that relay communication between the data providing device and the plurality of destination devices. A plurality of distribution destination devices based on topology information indicating a relationship and a communication speed of each of a plurality of links for communicatively connecting two of the data providing device, the plurality of distribution destination devices, and the plurality of relay devices; Among them, the distribution destination devices having the same link with the lowest communication speed on the communication path with the data providing device are classified into the same group. Next, the computer specifies, among the distribution destination devices belonging to the group, the first distribution destination device having the highest communication speed of the directly connected link. Then, the computer transmits the data from the data providing device to the first distribution destination device in the group, and then transmits the data from the first distribution destination device to the second distribution destination device other than the first distribution destination device in the group. Create a data distribution schedule so that

  According to one aspect, it is possible to reduce the time for creating a distribution schedule that allows efficient data distribution.

FIG. 2 is a diagram illustrating a configuration example of a distribution schedule creation device according to the first embodiment. FIG. 9 illustrates an example of a system configuration according to a second embodiment. FIG. 14 is a diagram illustrating a configuration example of hardware of a distribution management system used in a second embodiment. It is a figure showing an example of the data transfer of the first stage of multistage distribution. FIG. 9 is a diagram illustrating an example of data transfer in the second and subsequent stages of multi-stage distribution. It is a figure showing an example of data distribution scheduling by a BL method. It is a figure explaining the point of interest for improving distribution schedule efficiency. It is a figure showing an example of determination of the distribution range between ES. It is a figure which shows the example of determination of a representative ES. It is a block diagram showing the function of a distribution management system. It is a figure showing an example of distribution demand information. It is a figure showing an example of distribution data information. FIG. 4 is a diagram illustrating an example of NW topology information. It is a figure showing an example of NW link information. It is a figure showing an example of relay node information. It is a figure showing an example of ES information. It is a figure showing an example of FS information. It is a figure showing an example of range information. It is a figure showing an example of NW link use information. It is a figure showing an example of distribution schedule information. It is a flowchart which shows an example of a data distribution management procedure. It is a flowchart which shows an example of the procedure of range information generation processing. It is a flowchart which shows an example of the procedure of the distribution range creation between ES. It is a flowchart which shows an example of the procedure of a delivery schedule generation process. It is a flowchart which shows an example of the procedure of the distribution schedule creation process in a range. FIG. 7 is a diagram illustrating an example of creating a sub-range for the entire NW. FIG. 8 is a first diagram illustrating an example of generation of range information. FIG. 14 is a diagram illustrating an example of creating a sub-range for ranges “Grp1”, “Grp2”, and “Grp3”. FIG. 11 is a second diagram illustrating an example of generating range information. FIG. 13 is a diagram illustrating an example of creating a sub-range for ranges “Grp13” and “Grp21”. FIG. 13 is a third diagram illustrating an example of generation of range information. FIG. 14 is a diagram illustrating an example of determining a distribution schedule for a range “Grp2”. FIG. 14 is a diagram illustrating an example of distribution schedule information after a distribution schedule has been created for the entire range of the NW. It is a figure showing a data distribution situation according to a distribution schedule for the whole NW range. FIG. 11 is a diagram illustrating an example of NW link usage information after a distribution schedule has been created for the entire NW. FIG. 14 is a diagram illustrating an example of distribution schedule information after a distribution schedule has been created for ranges “Grp1”, “Grp2”, and “Grp3”. It is a figure which shows the data distribution status according to the distribution schedule for ranges "Grp1", "Grp2", and "Grp3". It is a figure showing an example of NW link use information of NW link "L4". It is a figure showing an example of distribution schedule information after distribution schedule creation for a range "Grp13" and "Grp21". It is a figure showing a data distribution situation according to a distribution schedule for ranges “Grp13” and “Grp21”. It is a figure showing an example of a notice of a distribution schedule. FIG. 14 is a diagram illustrating an example of an NW repository to which the third embodiment is applied. 9 is a flowchart illustrating an example of a data distribution management procedure when distributing a plurality of data. FIG. 14 is a diagram illustrating an example of determining a distribution schedule for a range “Grp2” of data “dataB”. It is a figure showing the data distribution situation according to the distribution schedule for the range “Grp2” of data “dataB”. FIG. 14 is a diagram illustrating an example of distribution schedule information after a distribution schedule for a range “Grp21” is created. It is a figure showing the data distribution situation according to the distribution schedule for the range “Grp21” of data “dataB”. It is a figure showing an example of distribution schedule information before distribution schedule determination of data distribution information "d05". It is a figure showing an example of NW link use information of NW link “L4” before distribution schedule of data distribution information “d05”. It is a figure which shows the example of the distribution schedule information after the distribution schedule of the data distribution information "d05" is determined. It is a figure which shows the example of the NW link use information of the NW link "L4" after the distribution schedule of the data distribution information "d05" is determined.

Hereinafter, the present embodiment will be described with reference to the drawings. Note that each embodiment can be implemented by combining a plurality of embodiments within a consistent range.
[First Embodiment]
FIG. 1 is a diagram illustrating a configuration example of a distribution schedule creation device according to the first embodiment. The distribution schedule creation device 10 includes a storage unit 11 and a calculation unit 12.

  The storage unit 11 stores topology information 11a of a network that distributes the data 1a. The network includes a data providing device 1, distribution destination devices 2a to 2g, and relay devices 3a to 3g. The data providing device 1 is a computer that stores data 1a to be distributed. The distribution destination devices 2a to 2g are devices such as computers to which data is distributed. The relay device 3a is a network device that relays communication between the data providing device 1 and the plurality of distribution destination devices 2a to 2g. In addition, the network includes a plurality of links (identifiers “L1” to “L13”) that communicatively connect two of the data providing device 1, the plurality of distribution destination devices 2a to 2g, and the plurality of relay devices 3a to 3g. ) Is included. The topology information 11a includes a connection relationship of the data providing apparatus 1, the distribution destination apparatuses 2a to 2g, and the relay apparatuses 3a to 3g by links, and communication speeds of a plurality of links (identifiers “L1” to “L13”). included.

  The calculation unit 12 creates a distribution schedule 13 of the data 1a based on the topology information 11a. For example, the arithmetic unit 12 sets the destination devices sharing the same link (bottleneck link) having the lowest communication speed on the communication path with the data providing device 1 among the plurality of destination devices 2a to 2g to the same destination device. They are classified into groups 4a to 4c. In the example of FIG. 1, the bottleneck link in the communication with the data providing device 1 is the link with the identifier “L1” in each of the distribution destination devices 2a to 2c. Therefore, the distribution destination devices 2a to 2c are grouped into the same group 4a. In each of the distribution destination devices 2d and 2e, the bottleneck link in the communication with the data providing device 1 is the link with the identifier “L8”. Therefore, the distribution destination devices 2d and 2e are grouped into the same group 4b. In each of the distribution destination devices 2f and 2g, the bottleneck link in the communication with the data providing device 1 is the link with the identifier “L9”. Therefore, the distribution destination devices 2f and 2g are grouped into the same group 4c.

  Next, the operation unit 12 specifies the first destination devices 2a, 2e, and 2g having the highest communication speed of the directly connected link among the destination devices belonging to the same groups 4a to 4c. Then, the calculation unit 12 creates the distribution schedule 13 based on the information of the distribution destination devices belonging to each of the groups 4a to 4c and the information of the specified first distribution destination device.

  The created distribution schedule 13 first indicates that the data 1a is transmitted from the data providing device 1 to the first distribution destination devices 2a, 2e, and 2g in the group. Further, in the distribution schedule 13, the first distribution destination devices 2a, 2e, and 2g that have acquired the data are transmitted from the first distribution destination devices 2a, 2e, and 2g to the second distribution destination devices 2b, 2c, 2d, and 2f other than the first distribution destination devices 2a, 2e, and 2g. The transmission of the data 1a is indicated.

By creating the distribution schedule 13 of the data 1a in this manner, the distribution schedule 13 that enables efficient data distribution can be created in a short time.
That is, in the distribution schedule 13, data transmission from the data providing device 1 to the distribution destination devices in each of the groups 4a to 4c is performed only once. Then, data distribution via a link having a low communication speed, which becomes a bottleneck, is performed only once, and the efficiency of data distribution is improved. Further, among the distribution destination devices in the group, the distribution destination device having the highest communication speed of the directly connected link is determined as the first distribution destination device. Thus, when data is distributed within a group, data can be transmitted at high speed from the first distribution destination device serving as a distribution source, and data transmission can be performed in a short time even when there are a plurality of distribution destinations. be able to. As a result, the overall distribution time required to complete data distribution to all distribution destination devices is reduced.

  Moreover, the amount of calculation required to create the distribution schedule 13 is smaller than when all patterns in the distribution order are generated for all combinations of the distribution source and the distribution destination and the overall distribution time is compared. Therefore, creation of a distribution schedule can be performed in a short time.

  When distributing data within each of the groups 4a to 4c, an efficient distribution schedule can be created by classifying the devices other than the first distribution destination device, which is the data distribution source, into subgroups. In this case, when there are a plurality of second distribution destination devices belonging to the groups 4a to 4c, the calculation unit 12 determines that the communication speed on the communication path between the second distribution destination device and the first distribution destination device is the lowest. The second delivery destination devices sharing a common bottleneck link are classified into the same subgroup. Next, the calculation unit 12 specifies, as the third destination device, the second destination device having the highest communication speed of the directly connected link among the second destination devices belonging to the subgroup. Then, the calculation unit 12 creates the distribution schedule 13 in consideration of the subgroup.

  For example, the created distribution schedule first indicates that the data 1a is transmitted from the data providing device 1 to the first distribution destination devices in the groups 4a to 4c. The distribution schedule 13 indicates that the data 1a is transmitted from the first distribution destination device that has received the data 1a to the third distribution destination device in the subgroup. The distribution schedule 13 indicates that the data 1a is transmitted from the third distribution destination device to the fourth distribution destination devices other than the third distribution destination device in the subgroup.

  As described above, the sub-groups are provided in the groups 4a to 4c, and the data transfer within the groups is efficiently performed, so that the total transmission time, which is the time required for data distribution to all the distribution destination devices 2a to 2g, Can be further reduced. By creating further subgroups within the subgroup, a distribution schedule can be created very efficiently even in a large-scale network.

  When there are a plurality of groups 4a to 4c, the order of transmitting data to the first distribution destination devices in each group can be determined so that the overall transmission time is shortened. For example, the operation unit 12 preferentially performs the transmission of the data 1a to the first destination device in the group, as the communication speed of the bottleneck link common to the first destination devices belonging to each group is lower. Next, a distribution schedule 13 is created. That is, the arithmetic unit 12 selects the groups in order from the group having the lowest communication speed of the bottleneck link, and transfers unused communication resources of each link to the first distribution destination of the selected group so that data transmission can be started as soon as possible. Allocate to data transmission processing to the device. As a result, the data transmission to the first distribution destination device, which has a low communication speed and takes a long time for data transmission, can be started early, and an increase in the overall transmission time is suppressed.

  Furthermore, when there are a plurality of data to be distributed, the order of the data to be transmitted can be determined so that the overall transmission time is shortened. For example, the calculation unit 12 calculates the total communication amount for each of the plurality of data based on the size of the data and the number of distribution destination devices requesting distribution of the data. The total communication volume is a value indicating the total amount of data transmitted until data is distributed to all the distribution destination devices requesting data distribution. For example, a value obtained by multiplying the number of distribution destination devices requesting data distribution by the data size is the total communication amount. Then, the calculation unit 12 creates the distribution schedule 13 such that transmission of the data is performed with higher priority for data having a larger total communication amount. That is, the arithmetic unit 12 selects the data in descending order of the total communication amount, and allocates the unused communication resources of each link to the distribution of the selected data requesting the data so that the data transmission can be started as soon as possible. Assign to transmission processing to the destination device. As a result, data that has a larger total communication amount and takes longer to distribute data to all of the distribution destination devices that request the data is transmitted earlier, and it is possible to suppress an increase in the overall transmission time.

  In some cases, the data communication speed can be dynamically changed under the control of the data providing device 1 and the distribution destination devices 2a to 2g. In such a case, if even a small band remains on the link, the band can be used effectively and data distribution can be performed efficiently. For example, the calculation unit 12 determines, for each distribution destination device, the data providing device 1 or another distribution destination device that is a data distribution source, a period during which the distribution destination device receives data distribution, and a communication speed of distribution. In determining the communication speed, the arithmetic unit 12 calculates, for each link on the communication path to the data distribution source device, the remaining bandwidth obtained by subtracting the communication speed of data distribution using the link from the communication speed of the link. Is calculated. Then, the arithmetic unit 12 determines, for each of the plurality of time zones, a communication speed that matches the remaining bandwidth of the link having the least remaining bandwidth in the time zone. As described above, by effectively using the small remaining bandwidth, the overall transmission time can be reduced.

  The calculation unit 12 can be realized by, for example, a processor included in the distribution schedule creation device 10. The storage unit 11 can be realized by, for example, a memory or a storage device included in the distribution schedule creation device 10.

[Second embodiment]
Next, a second embodiment will be described. The second embodiment is a more specific process of the first embodiment.

  FIG. 2 is a diagram illustrating an example of a system configuration according to the second embodiment. The terminal device 31 used by the data provider is connected to the distribution management system 100. The distribution management system 100 is a computer that manages data distribution. The distribution management system 100 is connected to the FS 32. The FS 32 is connected to a plurality of ESs 50 to 59 via a plurality of relay nodes 41, 42, 43,. Terminal devices 60 to 69 used by the data user are connected to the ESs 50 to 59, respectively. In the example of FIG. 2, only one terminal device is connected to one ES, but a plurality of terminal devices can be connected to one ES.

  The data provider stores the data to be distributed in the FS 32 via the distribution management system 100 using the terminal device 31. The data user uses the terminal devices 60 to 69 to specify data to be used. When the data to be used is specified, a data distribution request is transmitted from the ESs 50 to 59 to the distribution management system 100, for example. The distribution management system 100 creates a distribution schedule for the data stored in the FS 32 based on the data distribution request. After creating the delivery schedule, the delivery management system 100 instructs the FS 32 or each of the ESs 50 to 59 to transfer data according to the delivery schedule.

  FIG. 3 is a diagram illustrating a configuration example of hardware of a distribution management system used in the second embodiment. The distribution management system 100 is entirely controlled by a processor 101. The processor 102 is connected to the memory 102 and a plurality of peripheral devices via a bus 109. Processor 101 may be a multiprocessor. The processor 101 is, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or a DSP (Digital Signal Processor). At least a part of the function realized by the processor 101 executing the program may be realized by an electronic circuit such as an ASIC (Application Specific Integrated Circuit) or a PLD (Programmable Logic Device).

  The memory 102 is used as a main storage device of the distribution management system 100. The memory 102 temporarily stores at least a part of an OS (Operating System) program and an application program to be executed by the processor 101. Further, the memory 102 stores various data required for processing by the processor 101. As the memory 102, for example, a volatile semiconductor storage device such as a RAM (Random Access Memory) is used.

  The peripheral devices connected to the bus 109 include a storage device 103, a graphic processing device 104, an input interface 105, an optical drive device 106, a device connection interface 107, and a network interface 108.

  The storage device 103 electrically or magnetically writes and reads data to and from a built-in storage medium. The storage device 103 is used as an auxiliary storage device of a computer. The storage device 103 stores OS programs, application programs, and various data. As the storage device 103, for example, a hard disk drive (HDD) or a solid state drive (SSD) can be used.

  The monitor 21 is connected to the graphic processing device 104. The graphic processing device 104 displays an image on the screen of the monitor 21 according to a command from the processor 101. Examples of the monitor 21 include a display device using a CRT (Cathode Ray Tube) and a liquid crystal display device.

  A keyboard 22 and a mouse 23 are connected to the input interface 105. The input interface 105 transmits a signal transmitted from the keyboard 22 or the mouse 23 to the processor 101. The mouse 23 is an example of a pointing device, and another pointing device can be used. Other pointing devices include touch panels, tablets, touchpads, trackballs, and the like.

  The optical drive device 106 reads data recorded on the optical disk 24 using laser light or the like. The optical disk 24 is a portable recording medium on which data is recorded so as to be readable by light reflection. The optical disc 24 includes a DVD (Digital Versatile Disc), a DVD-RAM, a CD-ROM (Compact Disc Read Only Memory), and a CD-R (Recordable) / RW (ReWritable).

  The device connection interface 107 is a communication interface for connecting peripheral devices to the distribution management system 100. For example, the memory device 25 and the memory reader / writer 26 can be connected to the device connection interface 107. The memory device 25 is a recording medium having a function of communicating with the device connection interface 107. The memory reader / writer 26 is a device that writes data to the memory card 27 or reads data from the memory card 27. The memory card 27 is a card-type recording medium.

  The network interface 108 is connected to the network 20. The network interface 108 transmits and receives data to and from another computer or communication device via the network 20.

  With the hardware configuration described above, the processing functions of the second embodiment can be realized. The terminal devices 31, 60 to 69, the FS 32, the relay nodes 41, 42, 43,..., And the ESs 50 to 59 can also be realized by the same hardware as the distribution management system 100. Further, the distribution schedule creation device 10 shown in the first embodiment can be realized by the same hardware as the distribution management system 100 shown in FIG.

  The distribution management system 100 realizes the processing functions of the second embodiment, for example, by executing a program recorded on a computer-readable recording medium. The program describing the processing content to be executed by the distribution management system 100 can be recorded on various recording media. For example, a program to be executed by the distribution management system 100 can be stored in the storage device 103. The processor 101 loads at least a part of the program in the storage device 103 into the memory 102 and executes the program. Further, the program to be executed by the distribution management system 100 may be recorded on a portable recording medium such as the optical disk 24, the memory device 25, and the memory card 27. The program stored in the portable recording medium becomes executable after being installed in the storage device 103 under the control of the processor 101, for example. Further, the processor 101 can directly read out and execute the program from the portable recording medium.

Next, multi-stage distribution will be described with reference to FIGS.
FIG. 4 is a diagram illustrating an example of first-stage data transfer of multi-stage distribution. In the example of FIG. 4, it is assumed that “dataA” is delivered to ESs 50 to 55 and “dataB” is delivered to ESs 56 to 58. In the first-stage data transfer, each data is transmitted from the FS 32 to any of the ESs. In the example of FIG. 4, “dataA” is transmitted to the ES 50, and “dataB” is transmitted to the ES 56.

  FIG. 5 is a diagram illustrating an example of data transfer in the second and subsequent stages of multistage distribution. In the second and subsequent stages of data transmission, data transfer is performed between ESs. That is, in the second and subsequent stages, the ES that has received the data is the transmission source, and the data is transmitted to another ES that requests the same data. In the example of FIG. 5, "dataA" is transmitted from the ES 50 to the ESs 51, 52, and 54, and then transmitted from the ES 51 to the ES 53 and from the ES 52 to the ES 55. “DataB” is transmitted from the ES 56 to the ESs 57 to 58.

  Creating a distribution schedule for efficiently performing such multi-stage distribution is a combination optimization problem that determines a pairing of a distribution source and a distribution destination and a distribution order of each pair. At this time, if all the patterns are calculated, a combination explosion occurs, and the calculation cannot be completed within a practical time.

The calculation amount in the case of calculating all patterns is O (N ⁵ ) in order notation. The details are as follows.
[Step A] Calculation of distribution route: Calculation of a distribution route to all ESs with FS and all ESs as starting points (distribution sources).
[Step B] Determining a delivery schedule: Using the calculated route, calculating a delivery time in each combination pattern of a delivery source, a delivery destination, and a delivery order, and selecting a combination that minimizes the overall delivery time.

When the number of ESs is N (N is an integer of 1 or more), the calculation amount of each process can be estimated as follows.
・ Computational complexity of step A: O (N ² logN)
When the Dijkstra method is used, the calculation amount of the route calculation from one start point to all ESs is O (NlogN). In order to calculate a route starting from FS and all ESs, the calculation amount is O (N ² logN).

Calculation of Step B: O (N ⁵⁾
Distribution source, the number of combination of the delivery destination of the pair is N ^2. The distribution order for distributing each pair is N ² steps at the maximum when distributing one pair at a time. Since the timing for distributing each pair is at some point in the N ² steps, the total number of combinations is (N ² ) ² = N ⁴ . In each combination, for N data distributions, whether or not data transmission is performed beyond the bandwidth of the network (NW) link on the distribution route is checked, so that the calculation amount is O (N ⁵ ).

  When the distribution times for all the combinations are calculated in this way, the amount of calculation rapidly increases as the number of ESs (N) increases. Therefore, it is conceivable to create a distribution schedule that enables efficient data distribution by performing an approximate calculation. For example, data delivery scheduling can be considered a type of two-dimensional packing problem.

  The two-dimensional packing problem is a combination optimization problem that calculates how to pack each rectangle on a plane (combination of the arrangement of rectangles) when a given rectangle is placed on a given plane. For the two-dimensional packing problem, a solution (approximate solution) such as the bottom-left method (BL method) has been devised. The BL method is a method of repeatedly filling a rectangle as low and left as possible.

When the BL method is applied to the data distribution scheduling problem, the problem is how to calculate how to fill a rectangle corresponding to the data distribution amount in a space corresponding to the bandwidth (resource) of the NW link.
FIG. 6 is a diagram illustrating an example of data distribution scheduling by the BL method. In the BL method, a band and a time required for data transmission to each of the ESs 50 to 53 are represented by rectangles. For example, the length of the rectangle in the vertical direction indicates the used band, and the length in the horizontal direction indicates the distribution time. In this case, the rectangles are arranged so that the rectangles are packed as far down and to the left as possible in a space with the time on the horizontal axis and the band on the vertical axis. In this way, an efficient distribution schedule can be created.

  However, the BL method shown in FIG. 6 assumes only that the source is FS32. When multi-stage distribution is performed, there are a plurality of transmission sources, and it is difficult to simply create an efficient distribution schedule using only the BL method. Further, the two-dimensional packing problem targets one NW link, and cannot be applied to a plurality of NW links as they are.

Therefore, the second embodiment focuses on the following two points to improve the efficiency of the distribution schedule.
Characteristic A) Data transmission via an NW link with a small band requires a long transmission time.
Characteristic B) The earlier the data is received, the more data is transmitted to more ESs.

  FIG. 7 is a diagram for explaining points of interest for improving the distribution schedule efficiency. In the example of FIG. 7, the bandwidth of the NW link between the relay node 41 and the relay node 42 is 40 Mbps, which is narrower (the communication speed is lower) than the bandwidth of the other NW links. Therefore, data transmission through the NW link between the relay node 41 and the relay node 42 takes time (characteristic A). The ES 50 that first receives data from the FS 32 transmits data to the other three ESs 51, 52, and 54. In contrast, the ES 56 transmits data to the other two devices, and the ESs 51 and 52 transmit data to the other device. As described above, an ES that has received data earlier has a larger number of data transmission times than other ESs (characteristic B). Thus, as the number of stages in the multistage distribution increases, the number of ESs that have already received data increases, and the number of ESs that have not received data decreases. As a result, as the data reception timing is later, the number of ESs for which data is to be distributed thereafter becomes smaller, and the number of data transmissions becomes smaller.

  The distribution management system 100 according to the second embodiment, by taking into account the above two points of interest, appropriately performs grouping of ES groups that perform inter-ES distribution, thereby increasing the efficiency of data transfer. Hereinafter, the grouped ES group is referred to as an ES distribution range (or simply, “range”). The inter-ES distribution is performed only within the inter-ES distribution range. The distribution range between ESs is an example of the group or the subgroup described in the first embodiment. In the second embodiment, the entire network is regarded as one range (large group), and another range generated within one range is referred to as a sub-range with respect to the range of the generation source.

  FIG. 8 is a diagram illustrating an example of determining the distribution range between ESs. The NW topology shown in FIG. 8 includes one FS, eight relay nodes, and ten ESs. The identifier “FS01” is assigned to the FS. This FS is hereinafter referred to as FS “FS01”. The identifiers “A” to “H” are assigned to the eight relay nodes. These relay nodes are hereinafter referred to as relay nodes “A” to “H”. The identifiers of “ES01” to “ES10” are respectively assigned to the ten ESs. Hereinafter, these ESs are referred to as ESs “ES01” to “ES10”.

  The distribution management system 100 narrows the range of data transmission between ESs in order to shorten the calculation time. Then, the distribution management system 100 reduces the number of distribution destination candidates in each range by decomposing the data into the narrowed-down range and performing a distribution schedule creation process.

  At the time of narrowing down, the distribution management system 100 focuses on the distribution route from the transmission source to each destination in order to shorten the overall transmission time, and the bottleneck link (BL: NW link having the minimum bandwidth in the distribution route) is set. The same ES is narrowed down as the range of distribution between ESs. That is, paying attention to the characteristic A, the distribution management system 100 shortens each data transmission time by minimizing data transmission via the NW link with a narrow band.

  The example of FIG. 8 is an example in which the range is determined based on the BL when data is distributed to the FS “FS01” to each ES “ES01” to “ES10”. In FIG. 8, the FS “FS01” and the ESs “ES01” to “ES10” are connected via the relay nodes “A” to “H”. The devices are connected by NW links with identifiers “L0” to “L17” (hereinafter, referred to as NW links “L0” to “L17”). In FIG. 8, the bandwidth (bps) of the NW link is shown along a line segment indicating each NW link.

  Here, the BL between the FS “FS01” and each of the ESs “ES01” to “ES10” will be considered. In each of the ESs “ES01” to “ES5”, the NW link “L1” is BL. Therefore, the ESs “ES01” to “ES5” are grouped into one range. In the ESs “ES06” to “ES08”, the NW link “L10” is BL. Therefore, the ESs “ES06” to “ES08” are grouped into one range. In the ESs “ES09” to “ES10”, the NW link “L15” is BL. Therefore, the ESs “ES09” to “ES10” are grouped into one range.

After determining the range, the distribution management system 100 determines the ES (representative ES) that receives data first in each range in order to reduce the data transmission time in each range of the distribution between ESs.
FIG. 9 is a diagram illustrating an example of determining a representative ES. The distribution management system 100 focuses on NW links directly connected to each of the ESs “ES01” to “ES10”, and selects an ES having the maximum bandwidth of the directly connected NW link as a representative ES. That is, paying attention to the characteristic B, the distribution management system 100 preferentially sends data to the representative ES whose throughput at the time of transmission is large, thereby improving the efficiency of the distribution between the ESs. In the example of FIG. 9, the ESs “ES01”, “ES08”, and “ES10” are the representative ESs in the range to which each belongs.

  The distribution management system 100 also generates a sub-range from the representative ES using the bottleneck link of the distribution route to another ES in the same range within each range, and selects the representative ES in the sub-range. As a result, the delivery of the second and subsequent stages of the multistage delivery can be performed efficiently.

  For example, within the range “1”, the ES “ES01” that has received data from the FS “FS01” transmits data to the ESs “ES02” to “ES04”. Next, the ES “ES04” transmits data to the ES “ES05”. Within the range “2”, the ES “ES08” that has received data from the FS “FS01” transmits data to the ES “ES06”. Next, the ES “ES06” transmits data to the ES “ES07”. Within the range “3”, the ES “ES10” that has received data from the FS “FS01” transmits data to the ES “ES09”.

The distribution management system 100 creates a distribution schedule so that data transmission is performed in such an order, so that efficient data distribution is performed.
FIG. 10 is a block diagram illustrating functions of the distribution management system. The distribution management system 100 includes a distribution request information storage unit 111, a distribution data information storage unit 112, a NW topology information storage unit 113, a NW link information storage unit 114, a relay node information storage unit 115, an ES information storage unit 116, and an FS information storage. Unit 117, ES distribution range creation management unit 121, ES distribution range creation unit 122, range information storage unit 123, distribution schedule creation management unit 131, distribution schedule creation unit 132, NW link use information storage unit 133, distribution schedule information It has a storage unit 134 and a distribution schedule notification unit 141.

  The distribution request information storage unit 111 stores information (distribution request information) relating to a distribution request for instructing distribution to the ES. The distribution data information storage unit 112 stores information on distribution data (distribution data information). The NW topology information storage unit 113 stores information indicating a network configuration (NW topology information). The NW link information storage unit 114 stores information indicating the bandwidth of each NW link (NW link information). The relay node information storage unit 115 stores information (relay node information) indicating an NW link connected to the relay node. The ES information storage unit 116 stores information (ES information) indicating the NW link connected to the ES. The FS information storage unit 117 stores information (FS information) indicating an NW link connected to the FS.

The inter-ES distribution range creation management unit 121 collects information used to create the inter-ES distribution range, and issues an inter-ES distribution range creation instruction to the inter-ES distribution range creation unit 122 based on the collected information.
The inter-ES distribution range creation unit 122 creates an inter-ES distribution range according to an instruction from the inter-ES distribution range creation management unit 121. The inter-ES distribution range creation unit 122 stores the created inter-ES distribution range in the range information storage unit 123.

The range information storage unit 123 stores information (range information) indicating an ES belonging to each ES distribution range, a representative ES, and the like.
The distribution schedule creation management unit 131 acquires the range information, and instructs the distribution schedule creation unit 132 to create a distribution schedule for each range. The distribution schedule creation management unit 131 integrates the distribution schedules for each range and stores them in the distribution schedule information storage unit 134.

  The distribution schedule creation unit 132 creates a distribution schedule for each range. For example, the distribution schedule creating unit 132 uses the NW link use information storage unit 133 to manage the NW link use schedule based on the created distribution schedule as NW link use information. Then, the distribution schedule creation unit 132 determines whether data can be distributed in an available bandwidth of the NW link while referring to the NW link usage information, and creates a distribution schedule that is as efficient as possible.

  The NW link usage information storage unit 133 stores information (NW link usage information) indicating the NW link usage schedule. The distribution schedule information storage unit 134 stores information (distribution schedule information) indicating the created distribution schedule.

  Note that the lines connecting the elements shown in FIG. 10 show a part of the communication path, and a communication path other than the illustrated communication path can be set. The function of each element illustrated in FIG. 10 can be realized by, for example, causing a computer to execute a program module corresponding to the element.

Hereinafter, information stored in each storage unit will be described in detail with reference to FIGS.
FIG. 11 is a diagram illustrating an example of the distribution request information. The distribution request information 111a includes, for example, columns of id, distribution destination, and request data. In the column of “id”, an identifier of the distribution request is set. In the field of the distribution destination, the identifier of the ES that is the distribution destination of the data specified by the distribution request is set. In the request data column, an identifier of data to be distributed is set. Based on such distribution request information 111a, it can be determined which data is distributed to which ES.

  FIG. 12 is a diagram illustrating an example of the distribution data information. The distribution data information 112a includes, for example, columns of id, size, and owning node. The identifier of the data to be distributed is set in the id column. In the size column, the data size is set, for example, in GB. In the column of “owning node”, an identifier of the FS holding the data is set. Based on such distribution data information 112a, the initial distribution source of data and the amount of data to be transmitted can be determined.

  FIG. 13 is a diagram illustrating an example of the NW topology information. The NW topology information 113a has, for example, columns of id, NW link, relay node, ES, and FS. The identifier of the network to be managed is set in the id column. In the column of NW link, identifiers of all NW links connecting between devices in the network are set. In the column of the relay node, identifiers of all the relay nodes in the network are set. In the ES column, identifiers of all ESs in the network are set. In the FS column, identifiers of all FSs in the network are set. The components of the network can be recognized based on such NW topology information 113a.

  FIG. 14 is a diagram illustrating an example of the NW link information. The NW link information 114a includes, for example, columns of “id” and “band”. In the column of id, an identifier of the NW link is set. In the band column, the communication speed (band) of the NW link is set, for example, in Mbps. Based on such NW link information, the bandwidth of each NW link is known.

  FIG. 15 is a diagram illustrating an example of the relay node information. In the relay node information 115a, for example, a column of “id” and a connection link are provided. In the column of “id”, an identifier of the relay node is set. In the column of connection link, an identifier of the NW link directly connected to the relay node is set. Based on such relay node information 115a, it is possible to determine which NW link each relay node relays data.

  FIG. 16 is a diagram illustrating an example of the ES information. The ES information 116a has, for example, columns of “id” and “connection link”. The identifier of the ES is set in the id column. In the column of connection link, the identifier of the NW link directly connected to the ES is set. Based on such ES information 116a, it can be determined which NW link each ES transmits and receives data. By combining the ES information 116a and the NW link information 114a, the bandwidth of the NW link directly connected to the ES can be determined.

  FIG. 17 is a diagram illustrating an example of the FS information. The FS information 117a has, for example, a column of id and a connection link. An FS identifier is set in the id column. In the column of connection link, an identifier of the NW link directly connected to the FS is set. Based on such FS information 117a, it can be determined which NW link each FS transmits data.

  FIG. 18 is a diagram illustrating an example of the range information. The range information 123a includes, for example, columns of a range id, a belonging ES, a bottleneck link (BL), a representative ES, and a sub-range. In the column of the range id, an identifier of the range indicating the entire network or the distribution range between ESs is set. In the column of belonging ES, an identifier of an ES belonging to the range is set. In the column of bottleneck link, an identifier of an NW link that is a BL common to ESs belonging to a range is set. In the column of representative ES, the identifier of the representative ES in the range is set. When the entire network is in the range, the FS is the representative ES. In the sub-range column, an identifier of a range (sub-range) obtained by further grouping ESs belonging to the range is set. Based on such range information 123a, an appropriate distribution order can be determined.

  FIG. 19 is a diagram illustrating an example of the NW link usage information. In the example of FIG. 19, NW link use information 133a, 133b, 133c,... Is provided for each NW link. Each of the NW link usage information 133a, 133b, 133c,... Has a column of a period start time, a period end time, a used band, and a remaining band. In the column of period start time, a start time of a period in which the bandwidth used by the NW link is constant is set. In the period end time column, the end time of a period in which the bandwidth used by the NW link is constant is set. The used bandwidth (communication amount per unit time) is set in the used bandwidth column in a period in which the used bandwidth of the NW link is constant. In the column of remaining bandwidth, a value (remaining bandwidth) obtained by subtracting the used bandwidth from the bandwidth indicating the maximum communication speed of the NW link is set. Based on such NW link usage information 133a, 133b, 133c,..., The remaining bandwidth of each NW link based on the previously created distribution schedule can be determined.

  FIG. 20 is a diagram illustrating an example of the distribution schedule information. The distribution schedule information 134a includes, for example, a record for each data distribution information. The distribution schedule information 134a includes, for example, columns of id, distribution data, distribution source, distribution destination, route, assumed TP, required time, scheduled start time, scheduled end time, and preceding data distribution. In the column of “id”, identifiers of data distribution information (“d00”, “d01”,...) are set. Hereinafter, each data distribution information having the identifier “d00”, “d01”,... Is referred to as data distribution information “d00”, “d01”,.

  In the distribution data column, an identifier of data to be distributed is set. The FS or ES identifier for transmitting the data to be distributed is set in the distribution source column. In the field of the distribution destination, the identifier of the ES that receives the data to be distributed is set. In the column of the route, the identifier of the NW link through which the data is distributed is set. In the column of assumed TP, a communication speed (assumed throughput: TP) assumed at the time of data distribution is set. As the assumed throughput, for example, a value indicating the band of the narrowest band on the data distribution route is set. In the required time column, the time required for data distribution is set. The required time is a value obtained by dividing the size of data to be distributed by the assumed throughput. In the column of scheduled start time, a scheduled time at which data distribution is to be started is set. In the column of the scheduled end time, a scheduled time at which the data distribution is to be ended is set. In the column of preceding data distribution, an identifier of data distribution information required to be completed before the corresponding data distribution is performed is set.

  An efficient distribution schedule is created by the distribution management system 100 configured as described above, and the distribution schedule is transmitted to the FS “FS01” or the ESs “ES01” to “ES10”. Hereinafter, the processing executed by the distribution management system 100 will be described with reference to FIGS.

  FIG. 21 is a flowchart illustrating an example of the data distribution management procedure. Hereinafter, the processing illustrated in FIG. 21 will be described along the step numbers. The following processing is executed, for example, when a data distribution instruction is input via the terminal device 31 by the data provider. Further, the processing can be started at a preset date and time.

  [Step S101] The inter-ES distribution range creation management unit 121 and the inter-ES distribution range creation unit 122 cooperate to generate the range information 123a (see FIG. 18). Details of the range information generation processing will be described later (see FIG. 22).

  [Step S102] The delivery schedule creation management unit 131 and the delivery schedule creation unit 132 cooperate to generate delivery schedule information 134a (see FIG. 20). Details of the distribution schedule generation processing will be described later (see FIG. 24).

  [Step S103] The distribution schedule notification unit 141 transmits a data distribution schedule to each ES indicated in the distribution schedule information to each ES. Based on this notification, each ES requests data from the FS “FS01” or the ES of the distribution source at the scheduled start time indicated in the distribution schedule information, and receives the data distribution.

Next, the procedure of the range information generation processing will be described.
FIG. 22 is a flowchart illustrating an example of the procedure of the range information generation process. Hereinafter, the processing illustrated in FIG. 22 will be described along with step numbers.

[Step S111] The inter-ES distribution range creation management unit 121 reads the distribution request information 111a from the distribution request information storage unit 111.
[Step S112] The inter-ES distribution range creation management unit 121 reads information on the NW topology. For example, the inter-ES distribution range creation management unit 121 reads the NW topology information 113a from the NW topology information storage unit 113. Next, the ES distribution range creation management unit 121 reads the NW link information 114a from the NW link information storage unit 114. Next, the inter-ES distribution range creation management unit 121 reads the relay node information 115a from the relay node information storage unit 115. Next, the inter-ES distribution range creation management unit 121 reads the ES information 116a from the ES information storage unit 116. Next, the inter-ES distribution range creation management unit 121 reads the FS information 117a from the FS information storage unit 117.

  [Step S113] The inter-ES distribution range creation management unit 121 sets the entire NW as an inter-ES range creation target. For example, the inter-ES distribution range creation management unit 121 adds a record of the range id “all” to the range information 123a. At this time, the inter-ES distribution range creation management unit 121 sets the identifier of the ES set as the distribution destination of the data to be distributed in the distribution request information 111a among all the ESs “ES01” to “ES10” in the NW. Set as belonging ES of the added record. Further, the inter-ES distribution range creation management unit 121 sets the identifier of the FS “FS01” as the representative ES of the added record.

  [Step S114] The inter-ES distribution range creation management unit 121 instructs the inter-ES distribution range creation unit 122 to create an inter-ES distribution range within the range of the inter-ES distribution range creation target. At this time, the inter-ES distribution range creation management unit 121 transmits the information on the NW topology read in step S112 to the inter-ES distribution range creation unit 122. The inter-ES distribution range creation unit 122 creates an inter-ES distribution range according to the instruction. Details of the ES distribution range creation process will be described later (see FIG. 23).

  [Step S115] The inter-ES distribution range creation management unit 121 refers to the range information 123a, and determines whether there is a range in which the number of ESs to which the sub-range belongs is 2 or more, among the ranges for which the sub-range is not set. If there is such a range, the process proceeds to step S116. If there is no corresponding range, the range information generation processing ends.

  [Step S116] The inter-ES distribution range creation management unit 121 sets one of the inter-ES distribution ranges to which the number of ESs belongs to among the ranges for which no sub-ranges are set as the inter-ES distribution range creation target, and Proceeds to step S114.

Next, details of the inter-ES distribution range creation processing will be described.
FIG. 23 is a flowchart illustrating an example of a procedure of an ES distribution range creation process. Hereinafter, the processing illustrated in FIG. 23 will be described along the step numbers. The inter-ES distribution range creation processing is executed, for example, in response to a call from the inter-ES distribution range creation management unit 121 to the inter-ES distribution range creation processing.

  [Step S121] The inter-ES distribution range creation unit 122 calculates a route from a representative ES in the target range to each ES in the distribution range. For example, the inter-ES distribution range creation unit 122 reproduces the NW topology shown in FIG. 8 based on the NW topology information 113a, the relay node information 115a, the ES information 116a, and the FS information 117a. Then, the inter-ES distribution range creation unit 122 creates a list of NW links on the route from the representative ES to the corresponding ES for each ES other than the representative ES.

  [Step S122] The inter-ES distribution range creation unit 122 determines the BL on the route for each ES other than the representative ES based on the NW link information 114a. For example, the bandwidth of each NW link on the route of each ES is compared, and the NW link with the narrowest bandwidth is BL. Next, the inter-ES distribution range creation unit 122 groups, among the ESs other than the representative ES, the ESs with the same BL into the same inter-ES distribution range (sub-range).

  The ES distribution range creation unit 122 sets the identifier of the sub-range generated by the grouping in the sub-range column of the target range record in the range information 123a. Further, the inter-ES distribution range creation unit 122 adds a record corresponding to the generated sub-range to the range information 123a. In the column of the range id of the added record, the identifier of the sub-range is set. In the column of the belonging ES of the added record, an identifier of an ES having a common BL within the target range is set. In the column of the bottleneck link of the added record, the identifier of the NW link which is the BL common to the ESs belonging to the sub-range is set.

[Step S123] The inter-ES distribution range creation unit 122 selects one of the sub-ranges generated in Step S122.
[Step S124] Based on the ES information 116a, the inter-ES distribution range creation unit 122 extracts, for each ES in the selected sub-range, an NW link directly connected to the ES. In addition, the ES distribution range creation unit 122 extracts the extracted NW link band based on the NW link information 114a.

  [Step S125] The inter-ES distribution range creation unit 122 selects, from among the ESs belonging to the selected sub-range, the ES having the maximum bandwidth of the directly connected NW link as the representative ES. The ES distribution range creation unit 122 sets the identifier of the selected ES in the column of the representative ES of the record of the sub-range in the range information 123a.

  [Step S126] The inter-ES distribution range creation unit 122 determines whether a representative ES has been selected for all of the sub-ranges generated in step S122. If the representative ESs in all the sub-ranges have been selected, the processing for creating a distribution range between ESs ends. If there is a sub-range for which the representative ES has not been selected, the process proceeds to step S123.

Thus, the range information 123a as shown in FIG. 18 is generated. After the generation of the range information 123a, a distribution schedule generation process is performed.
FIG. 24 is a flowchart illustrating an example of a procedure of a distribution schedule generation process. Hereinafter, the processing illustrated in FIG. 24 will be described along with step numbers.

[Step S131] The distribution schedule creation management unit 131 sets the entire NW as a distribution schedule creation target.
[Step S132] The distribution schedule creation management unit 131 instructs the distribution schedule creation unit 132 to create a distribution schedule within the range of the distribution schedule creation target. Then, the distribution schedule creation unit 132 executes the in-range distribution schedule creation processing, and creates a data distribution schedule from the representative ES in the range to be created to the representative ES in a sub-range of the range. Details of the in-range distribution schedule creation processing will be described later (see FIG. 25).

  [Step S133] The distribution schedule creation management unit 131 determines whether there is a range in which a distribution schedule has not been created in the range registered in the range information 123a. If there is such a range, the process proceeds to step S134. If the creation of the distribution schedule has been completed for all the ranges, the distribution schedule generation processing ends.

  [Step S134] The distribution schedule creation management unit 131 sets, as a distribution schedule creation target, one of the ranges for which the distribution schedule with the representative ES as the distribution destination has been created and for which the distribution schedule within the range has not been created. Set. Thereafter, the process proceeds to step S132.

Next, the in-range distribution schedule creation processing will be described in detail.
FIG. 25 is a flowchart illustrating an example of a procedure of the in-range distribution schedule creation processing. Hereinafter, the processing illustrated in FIG. 25 will be described along with step numbers. Note that the range for which the distribution schedule is created is referred to as a target range.

  [Step S141] The distribution schedule creation unit 132 sorts the sub-ranges created in the target range in ascending order of BL. For example, the distribution schedule creation unit 132 refers to the record of the target range in the range information 123a, and extracts a sub-range of the target range. Next, the distribution schedule creation unit 132 acquires a bottleneck link from each record in which the identifier of the extracted sub-range is set in the range id. Next, the distribution schedule creation unit 132 acquires the BL bandwidth of each sub-range from the NW link information 114a. Then, the distribution schedule creating unit 132 sorts the sub-ranges in ascending order of the value of the band of BL in the range.

  [Step S142] The distribution schedule creation unit 132 acquires the representative ESs of the sub-ranges in the sort order (in order of smaller bandwidth). For example, the distribution schedule creation unit 132 specifies the highest sub-range in the sort order among the sub-ranges in which the representative ES has not been acquired, which is the sub-range of the target range. Then, the distribution schedule creating unit 132 acquires the representative ES of the record corresponding to the specified sub-range from the range information 123a.

  [Step S143] The distribution schedule creation unit 132 creates data distribution information from the representative ES in the target range to the representative ES in the sub-range acquired in step S142. The data distribution information includes the identifier of the data to be distributed, the identifier of the distribution source (representative ES of the target range), the distribution destination (identifier of the sub-range representative ES), the route from the distribution source to the distribution destination, the assumed throughput, and the required Time is included. The distribution schedule creation unit 132 registers the created data distribution information as a new record in the distribution schedule information 134a.

  [Step S144] The distribution schedule creation unit 132 specifies a period during which all the NW links on the route can provide the assumed throughput. For example, the distribution schedule creating unit 132 refers to the NW link usage information of the NW links on the route, and specifies a period (distributable period) in which the assumed throughput can be provided in all the NW links on the route. In the NW link usage information, a period in which the remaining bandwidth is equal to or greater than the assumed throughput is a period in which the assumed throughput can be provided in the NW link (a period that can be provided). The period that can be provided in all NW links on the route is the distribution available period.

  [Step S145] The delivery schedule creation unit 132 determines the earliest time slot in the required time within the deliverable period obtained in step S144 as the delivery time slot. Then, the distribution schedule creation unit 132 converts the distribution time period of each NW link on the route and the bandwidth for the assumed throughput in that period from the representative ES of the target range to the representative ES of the acquired sub-range. Assign to transmission.

  [Step S146] The distribution schedule creation unit 132 adds the scheduled distribution start time and the scheduled distribution end time to the data distribution information. For example, the distribution schedule creation unit 132 sets the start time of the distribution time zone determined in step S145 as the scheduled distribution start time. The distribution schedule creation unit 132 sets the end time of the distribution time zone as the scheduled distribution end time. Then, the distribution schedule creation unit 132 sets the scheduled distribution start time and the scheduled distribution end time in the record added in step S144 to the distribution schedule information 134a.

  [Step S147] The distribution schedule creating unit 132 determines whether data distribution information for distributing data to the representative ES has been created for all sub-ranges in the target range. When the creation of the data distribution information is completed for all the sub-ranges, the in-range distribution schedule creation processing ends. If there is a sub-range for which data distribution information has not been created, the process proceeds to step S142.

  In this way, a distribution schedule is created. By performing data distribution according to the created distribution schedule, the frequency of use of the NW link with a narrow band is reduced, and the overall distribution time is shortened. Further, as compared with a case where all combinations of data distribution are generated, a distribution schedule can be easily created, and a distribution schedule can be quickly created.

  Hereinafter, in the NW topology shown in FIG. 8, specific data when the data of the identifier “dataA” (hereinafter, referred to as data “dataA”) is distributed from the FS “FS01” to the ESs “ES01” to “ES10” An example of distribution will be described. It is assumed that the size of the data “dataA” is 18 GB. First, with reference to FIGS. 26 to 31, a description will be given of a status of creating a distribution range between ESs.

  FIG. 26 is a diagram illustrating an example of creating a sub-range for the entire NW. When the entire range is targeted, grouping of ESs “ES01” to “ES10” starting from the FS “FS01” is performed. For example, a distribution route to each of the ESs “ES01” to “ES10” starting from the FS “FS01” is calculated, and ESs having the same NW link (BL) having the minimum bandwidth are grouped into the same sub-range. Is done.

  In the example of FIG. 26, ESs “ES01” to “ES05” are grouped in a range of an identifier “Grp1” (hereinafter, referred to as a range “Grp1”). In the ESs “ES01” to “ES05” belonging to the range “Grp1”, the NW link “L1” between the relay node “A” and the relay node “B” is BL.

  ESs “ES06” to “ES08” are grouped in a range of an identifier “Grp2” (hereinafter, referred to as a range “Grp2”). In the ESs “ES06” to “ES08” belonging to the range “Grp2”, the NW link “L10” between the relay node “C” and the relay node “F” is BL.

  ESs “ES09” to “ES10” are grouped in a range of an identifier “Grp3” (hereinafter, referred to as a range “Grp3”). In the ESs “ES09” to “ES10” belonging to the range “Grp3”, the NW link “L15” between the relay node “C” and the relay node “G” is BL.

  When a new range (sub-range) is generated, a representative ES is selected for each sub-range. In the selection of the representative ES, among the ESs included in each sub-range, the ES with the largest directly connected NW link is selected as the representative ES.

  For example, in the case of the range “Grp1” shown in FIG. 26, five ESs “ES01” to “ES05” are included. The band of the link “L4” directly connected to the ES “ES01” is “100 Mbps”. The bandwidth of the link “L5” directly connected to the ES “ES02” is “80 Mbps”. The band of the link “L6” directly connected to the ES “ES03” is “50 Mbps”. The band of the link “L7” directly connected to the ES “ES04” is “80 Mbps”. The bandwidth of the link “L8” directly connected to the ES “ES05” is “50 Mbps”.

Among the ESs “ES01” to “ES05”, the ES “ES01” has the largest band of the NW link directly connected. Therefore, the ES “ES01” is selected as the representative ES.
Similarly, the ES “ES08” is selected as the representative ES of the range “Grp2”. The ES “ES10” is selected as the representative ES of the range “Grp3”.

As described above, the range information 123a is updated in accordance with the creation result of the sub-range and the selection result of the representative ES.
FIG. 27 is a first diagram illustrating an example of generation of range information. In the example of FIG. 27, in the range information 123a, a record in which the entire NW is a range (range id "all") and records corresponding to the generated sub-ranges (ranges "Grp1,""Grp2,""Grp3") Has been added. Also, the ranges “Grp1”, “Grp2”, and “Grp3” generated as the sub-ranges are added to the sub-range column of the record of the entire NW.

  A plurality of ESs are included in the sub-range created by the processing shown in FIGS. Therefore, for each created sub-range, ESs belonging to the sub-range are grouped to create a sub-range. Then, the representative ES of the created sub-range is selected.

FIG. 28 is a diagram illustrating an example of creating a sub-range for the ranges “Grp1”, “Grp2”, and “Grp3”.
For the range “Grp1”, three sub-ranges with identifiers “Grp11”, “Grp12”, and “Grp13” (hereinafter, referred to as ranges “Grp11”, “Grp12”, and “Grp13”) are created. The representative ES in the range “Grp11” is the ES “ES02”. The representative ES in the range “Grp12” is the ES “ES03”. The representative ES in the range “Grp13” is the ES “ES04”.

  For the range “Grp2”, one sub-range having the identifier “Grp21” (hereinafter, referred to as a range “Grp21”) is created. The representative ES in the range “Grp21” is the ES “ES06”.

  For the range “Grp3”, one sub-range having the identifier “Grp31” (hereinafter, referred to as a range “Grp31”) is created. The representative ES in the range “Grp31” is the ES “ES09”.

  Then, information corresponding to the sub-ranges (ranges “Grp11”, “Grp12”, “Grp13”, “Grp21”, “Grp31”) as shown in FIG. 28 is added to the range information 123a.

  FIG. 29 is a second diagram illustrating an example of generation of range information. In the example of FIG. 29, records corresponding to the generated sub-ranges (ranges id “Grp11”, “Grp12”, “Grp13”, “Grp21”, “Grp31”) are added to the range information 123a. Further, the identifier of the generated sub-range is set in the sub-range column of the record of the generation source range.

  Of the sub-ranges created by the processing shown in FIGS. 28 and 29, each of the ranges “Grp13” and “Grp21” includes a plurality of ESs. Therefore, for these ranges, further sub-ranges are generated.

FIG. 30 is a diagram illustrating an example of creating a sub-range for the ranges “Grp13” and “Grp21”.
For the range “Grp13”, one sub-range having the identifier “Grp131” (hereinafter, referred to as a range “Grp131”) is created. The representative ES in the range “Grp131” is the ES “ES05”.

  For the range “Grp21”, one sub-range having the identifier “Grp211” (hereinafter, referred to as a range “Grp211”) is created. The representative ES in the range “Grp211” is the ES “ES07”.

Then, information corresponding to the sub-ranges (ranges “Grp131” and “Grp211”) as illustrated in FIG. 30 is added to the range information 123a.
FIG. 31 is a second diagram illustrating an example of generation of range information. In the example of FIG. 31, records corresponding to each of the generated sub-ranges (range id “Grp131”, “Grp211”) are added to the range information 123a. Further, the identifier of the generated sub-range is set in the sub-range column of the record of the generation source range.

  Of the records registered in the range information 123a shown in FIG. 31, the records to which the sub-ranges are not set belong to one ES. Accordingly, the range information generation processing ends, and the distribution schedule generation processing is performed based on the range information 123a shown in FIG.

  First, a distribution schedule is created for the entire range of the NW. In this case, first, the ranges “Grp1”, “Grp2”, and “Grp3”, which are sub-ranges for the entire NW, are sorted in ascending order of the BL band. Then, the ranges are “Grp2”, “Grp1”, and “Grp3” in that order. Next, the distribution schedule to the representative ES is determined in the sort order.

  FIG. 32 is a diagram illustrating an example of determining a distribution schedule for the range “Grp2”. The representative ES in the range “Grp2” is the ES “ES08”. Therefore, data distribution information “d00” indicating data distribution from FS “FS01” to ES “ES08” is added to the distribution schedule information 134a as one record. The data distribution time zone is determined based on the NW link use information 133a to 133d of the NW link indicated in the route of the data distribution information. That is, for all NW links (NW links “L0”, “L9”, “L10”, “L14”) in the route, the period (distributable period) in which the bandwidth for the assumed throughput can be provided for the required time is specified. Is done. In the example of FIG. 32, since data distribution is not allocated to any NW link, if the data distribution start time is from “00:00”, a bandwidth of 20 Mbps can be provided between “00:00 to 02:00”. It is. Therefore, “0:00 to 02:00” is determined as the distribution period, and the band allocation result of the specified period is reflected in the usage information of all NW links. Then, the scheduled start time “00:00” and the scheduled end time “02:00” are added to the data distribution information “d00”.

  For the remaining subranges (“Grp1” and “Grp2”), the distribution schedule from the FS “FS01” to the representative ES is determined in the sort order. Then, the distribution schedule information 134a is updated according to the determined schedule.

  FIG. 33 is a diagram illustrating an example of distribution schedule information after a distribution schedule has been created for the entire range of the NW. As shown in FIG. 33, the data distribution information from the FS “FS01” to the ESs “ES01” and “ES10” that are the representative ESs in each of the sub-ranges (“Grp1” and “Grp2”) is included in the distribution schedule information 134a. Has been added. According to such distribution schedule information 134a, data distribution is performed with the FS “FS01” as the distribution source.

  FIG. 34 is a diagram illustrating a data distribution status according to a distribution schedule for the entire range of the NW. As shown in FIG. 34, data is distributed from FS “FS01” to one ES in each sub-range.

When the distribution period of the data distribution information is determined, the NW link use information of the NW link on the route is updated.
FIG. 35 is a diagram illustrating an example of NW link usage information after creating a distribution schedule for the entire range of the NW. As shown in FIG. 35, the NW link use information 133a, 133b, 133e for each NW link includes the time of the used band and the remaining band of each NW link when data distribution is performed according to the decided distribution schedule. A transition is indicated. For example, on the NW link with the identifier “L0”, three data distributions are simultaneously started to be executed from time “00:00”, and a bandwidth of 75 Mbps is used. At time “01:20”, the data distribution to the ES “ES10” ends, and the bandwidth used drops to 45 Mbps. Further, at time “01:36”, the data distribution to the ES “ES01” ends, and the bandwidth used drops to 20 Mbps.

  Next, a distribution schedule for each of the sub-ranges (ranges “Grp1”, “Grp2”, “Grp3”) created from the entire distribution NW is determined. Then, the distribution schedule information 134a is updated according to the determined schedule.

  FIG. 36 is a diagram illustrating an example of distribution schedule information after a distribution schedule has been created for the ranges “Grp1”, “Grp2”, and “Grp3”. For example, for the range “Grp1”, a distribution schedule is created for sub-ranges (range “Grp11”, “Grp12”, “Grp13”) for the range “Grp1”, and data distribution information indicating the relevant distribution schedule is distributed schedule information. 134a. The preceding data distribution in the added data distribution information is distribution to the distribution source ES “ES01” (data distribution information “d00”). This indicates that the data distribution according to the data distribution information “d03”, “d04”, and “d05” cannot be started unless the data is distributed to the ES “ES01” that is the distribution source.

  FIG. 37 is a diagram illustrating a data distribution status according to a distribution schedule for the ranges “Grp1”, “Grp2”, and “Grp3”. As shown in FIG. 37, in each of the ranges “Grp1”, “Grp2”, and “Grp3”, one ES in each of the sub-ranges is changed from the representative ESs “ES01”, “ES08”, and “ES10”. Data is delivered.

  It should be noted that there are three data distribution destinations having the ES “ES01” as the distribution source, ie, the ESs “ES02” to “ES04”. Whether data distribution can be started to these three devices at the same time depends on the bandwidth used and the remaining bandwidth of the NW link “L4”, which are overlapping routes.

  FIG. 38 is a diagram illustrating an example of NW link usage information of the NW link “L4”. The time zone from time “00:00” to time “01:36” is used for data distribution based on the data distribution information “d00”. Based on the availability of the NW link “L4”, the data distribution information “d03” and “d04” can be simultaneously distributed from time “01:36”. However, the remaining bandwidth of the NW link with the identifier “L4” from time “01:36” is only 20 Mbps. This remaining bandwidth is less than the assumed TP (80 Mbps) in the data distribution information “d05”. Therefore, until the time “2:56” at which the remaining bandwidth of the NW link “L4” becomes 80 Mbps or more, the start of data distribution according to the data distribution information “d05” is waited.

  As shown in FIG. 36, for the ranges "Grp2" and "Grp3", data distribution information "d06" and "d07" indicating the distribution schedule to the internal sub-range are generated and added to the distribution schedule information 134a. You.

  Next, a distribution schedule is determined for each of the ranges “Grp11”, “Grp12”, “Grp13”, “Grp21”, and “Grp31” created from the ranges “Grp1”, “Grp2”, and “Grp3”. You. Note that among the ranges “Grp11”, “Grp12”, and “Grp13” created from the range “Grp1”, the processing is skipped for the ranges “Grp11” and “Grp12” because no sub-range has been created. Also, for each range “Grp31” created from the range “Grp3”, the process is skipped because no sub-range has been created. As a result, a distribution schedule for the ranges “Grp13” and “Grp21” is determined. Then, the distribution schedule information 134a is updated according to the determined schedule.

  FIG. 39 is a diagram illustrating an example of distribution schedule information after a distribution schedule has been created for the ranges “Grp13” and “Grp21”. For example, for the range “Grp13”, a distribution schedule is created for a subrange (range “Grp131”) for the range “Grp13”, and data distribution information “d08” indicating the distribution schedule is added to the distribution schedule information 134a. . In addition, for the range “Grp21”, a distribution schedule for a subrange (range “Grp211”) for the range “Grp21” is created, and data distribution information “d09” indicating the distribution schedule is added to the distribution schedule information 134a. .

  FIG. 40 is a diagram illustrating a data distribution status according to a distribution schedule for the ranges “Grp13” and “Grp21”. As shown in FIG. 40, in each of the ranges “Grp13” and “Grp21”, data is distributed from one of the representative ESs “ES04” and “ES06” to one ES in each of the sub-ranges. As a result, the data has been distributed to all ESs. That is, if data distribution is performed according to the distribution schedule information shown in FIG. 39, data distribution is performed in all ESs. As a result, the creation of the distribution schedule ends.

When the creation of the delivery schedule is completed, the delivery schedule notification unit 141 notifies each ES of the delivery schedule information indicating the created delivery schedule.
FIG. 41 is a diagram illustrating a notification example of a distribution schedule. The distribution schedule information 61, 62, 63,... Is transmitted from the distribution management system 100 to the ESs 51, 52, 53,. The transmitted delivery schedule information 61, 62, 63,... Includes, for example, “distribution source, delivery data, scheduled delivery start time, scheduled delivery end time”. Each of the ESs 51, 52, 53,... Starts acquiring data from the distribution source when its own scheduled distribution time comes based on the received distribution schedule information 61, 62, 63,.

  In this way, the distribution order of the distribution destination ES is determined so that the distribution that takes a long time to distribute (the data distribution via the NW link having a small bandwidth) is suppressed and the data distribution between the ESs proceeds quickly. Delivery time can be shortened.

Further, since there is no need to compare a plurality of distribution order patterns, the amount of calculation can be reduced. For example, when the number of ESs is N, the calculation amount of the process performed in one loop is O (NlogN). The calculation method of the calculation amount is as follows.
A) Calculation of distribution route: The calculation amount of the route from the start point to all the remaining ESs is O (NlogN) at the maximum.
B) Extraction of range: The amount of calculation is O (N) because the routes to each ES calculated in A) are checked one by one.
C) Identification of representative ES: In each group calculated in B), the band of the NW link directly connected to all the ESs to which the group belongs is confirmed. Since the calculation is performed for the number of ESs belonging to all the groups, the calculation amount is O (N) at the maximum.
D) Schedule calculation from start point to representative ES: For each representative ES calculated in C), the remaining bandwidth of the NW link on the route is confirmed. Since the calculation is performed for the ESs belonging to the group, the calculation amount is O (N) at the maximum.

The above processing is performed once for the entire NW and for the range where the number of ESs is 2 or more. Since the number of applicable ranges can be N at the maximum, the calculation amount is O (N ² logN) at worst. If the number of ESs cannot be in a range of 2 or more (in all ESs, the directly connected NW link is BL), the calculation amount of the calculation for the range is O (NlogN). This calculation amount is very small as compared with the calculation amount O (N ⁵ ) when solving the combination optimization problem of all patterns. That is, an efficient distribution schedule can be created with a small calculation load.

[Third Embodiment]
Next, a third embodiment will be described. The third embodiment is an example in the case of distributing a plurality of data.

FIG. 42 is a diagram illustrating an example of an NW repository to which the third embodiment is applied. When transmitting a plurality of data, priority is given to creating a distribution schedule in descending order of the total request data size. The total required data size can be calculated by the following formula.
Total request data size = Request data size × Number of ESs requesting each data To create a distribution schedule in descending order of the total request data size, data with a larger data size and a larger number of distributions must be completed before distribution is completed. This is because it takes time.

  In the example of FIG. 42, it is assumed that 18 GB of data “dataA” and 15 GB of data “dataB” are distributed. The number of ESs requesting the data “dataA” is 10, and the number of ESs requesting the data “dataB” is 2. In this case, the total required size of the data “dataA” is 18 GB × 10 ES = 180 GB, and the total required size of the data “dataB” is 15 GB × 2 ES = 30 GB. As a result, a distribution schedule for the data “dataA” is created with priority.

Next, a data distribution management procedure according to the third embodiment will be described.
FIG. 43 is a flowchart illustrating an example of a data distribution management procedure when distributing a plurality of data. Hereinafter, the processing illustrated in FIG. 43 will be described along the step numbers.

  [Step S201] The inter-ES distribution range creation management unit 121 and the inter-ES distribution range creation unit 122 cooperate to generate the range information 123a (see FIG. 18). Details of the range information generation process are as shown in FIG.

[Step S202] For each piece of data to be distributed, the distribution schedule creation management section 131 calculates the total traffic of the data.
[Step S203] The distribution schedule creation management unit 131 selects data to be scheduled in descending order of the total traffic.

  [Step S204] The delivery schedule creation management unit 131 and the delivery schedule creation unit 132 cooperate to generate delivery schedule information 134a (see FIG. 20) of the data selected as the delivery target. Details of the distribution schedule generation process are as shown in FIG.

  [Step S205] The distribution schedule creation management unit 131 determines whether all data has been selected as a scheduling target. If there is unselected data, the process proceeds to step S203. If all data has been selected, the process proceeds to step S206.

[Step S106] The distribution schedule notification unit 141 notifies each ES of the data distribution schedule to each ES indicated in the distribution schedule information.
By such a procedure, a distribution schedule is created for each of the plurality of data. In the example shown in FIG. 42, the distribution schedule of the data “dataA” is created first, and then the distribution schedule of the data “dataB” is created. The distribution schedule of the data “dataA” is as shown in FIG. Hereinafter, an example of creating a distribution schedule of the data “dataB” will be described. In the example of FIG. 42, the ESs “ES06” and “ES08” that have issued a request to receive the data “dataB” belong to the range “Grp2”. Therefore, first, a distribution schedule for the range “Grp2” is determined.

  FIG. 44 is a diagram illustrating an example of determining a distribution schedule for a range “Grp2” of data “dataB”. The representative ES in the range “Grp2” is the ES “ES08”. Therefore, data distribution information “d10” indicating data distribution from FS “FS01” to ES “ES08” is added to the distribution schedule information 134a as one record. The data distribution time zone is determined based on the NW link use information 133a to 133d of the NW link indicated in the route of the data distribution information. That is, for all NW links (NW links “L0”, “L9”, “L10”, “L14”) in the route, the period (distributable period) in which the bandwidth for the assumed throughput can be provided for the required time is specified. Is done. In the example of FIG. 44, a bandwidth of 20 Mbps can be provided after the time “02:00”. Therefore, “02: 00 to 03: 40” is determined as the distribution period, and the band allocation result of the specified period is reflected in the usage information of all NW links. Then, the scheduled start time “02:00” and the scheduled end time “03:40” are added to the data distribution information “d10”.

  FIG. 45 is a diagram illustrating a data distribution status according to a distribution schedule for the range “Grp2” of the data “dataB”. As shown in FIG. 45, data “dataB” is distributed from the FS “FS01” to the ES “ES08” that is the representative ES in the range “Grp2”.

Similarly, a distribution schedule for a sub-range (range “Grp21”) for range “Grp2” is determined. Then, the distribution schedule information 134a is updated.
FIG. 46 is a diagram illustrating an example of the distribution schedule information after the distribution schedule for the range “Grp21” is created. As shown in FIG. 46, in addition to the data distribution information “d10” shown in FIG. 44, the data distribution information “ES08” from the ES “ES08” to the ES “ES06” that is the representative ES of the sub-range (range “Grp21”) “d11” has been added to the distribution schedule information 134a. According to such distribution schedule information 134a, data distribution is performed with the FS “FS01” as the distribution source.

  FIG. 47 is a diagram illustrating a data distribution status according to a distribution schedule for the range “Grp21” of the data “dataB”. As shown in FIG. 47, data “dataB” is distributed from ES “ES08” to ES “ES06” which is a representative ES within the range “Grp21”. As a result, the data “dataB” is distributed to all ESs that request the distribution of the data “dataB”, and the distribution schedule creation processing ends.

  By creating a distribution schedule so that data having a large total communication volume is transmitted with priority as described above, it is possible to execute data distribution that takes a long time to complete the distribution in advance, and to perform overall data distribution. The end time can be advanced.

[Fourth Embodiment]
Next, a fourth embodiment will be described. In the fourth embodiment, the data transmission rate can be dynamically changed to improve the efficiency of data distribution.

  In the example shown in FIG. 38 of the second embodiment, the distribution schedule indicated by the data distribution information “d05” is such that the remaining bandwidth (20 Mbps) from the time “01:36” of the NW link “L4” is the data It is less than the assumed throughput (80 Mbps) of the distribution information “d05”. In the second embodiment, in such a case, data distribution based on the data distribution information "d05" is waited until the remaining bandwidth of the NW link "L4" becomes 80 Mbps or more. However, in the case of controlling the transmission rate (throughput) of each distribution, further shortening the distribution time can be expected by performing the distribution earlier using the free bandwidth even under the assumed TP.

  Here, it is assumed that a distribution schedule has been created for the entire range of the NW, and a distribution schedule for the ranges “Grp1”, “Grp2”, and “Grp3” will be created. The distribution schedule determined at this time is as shown in FIG. In such a state, a distribution schedule for the ranges “Grp1”, “Grp2”, and “Grp3” as shown in FIG. 47 is created. In this case, the data distribution information “d03” and “d04” are created in the same manner as in the second embodiment. After that, at the stage of creating the data delivery information “d05”, a delivery schedule is created on the assumption that the data transfer rate is dynamically changed.

  FIG. 48 is a diagram illustrating an example of distribution schedule information before the distribution schedule of the data distribution information “d05” is determined. As shown in FIG. 48, when changing the data transfer rate, the assumed TP and the required time are not determined until the distribution schedule is determined.

The distribution schedule of the data distribution information “d05” depends on the use status of the NW link “L4” at the time of determining the distribution schedule of the data distribution information “d03” and “d04”.
FIG. 49 is a diagram illustrating an example of the NW link use information of the NW link “L4” before the distribution schedule of the data distribution information “d05” is determined. As shown in FIG. 49, the bandwidth of the NW link “L4” is used for the data distribution according to the data distribution information “d00”, “d03”, and “d04”.

  The preceding data distribution for the data distribution information “d05” is the data distribution information “d01”. Therefore, after the scheduled end time “01:36” of the data distribution information “d01”, data distribution according to the data distribution information “d05” becomes possible.

  Therefore, in the time zone “01:36 to 02:24” (48 minutes), data distribution according to the data distribution information “d05” can be performed with the remaining bandwidth (20 Mbps) as the assumed TP. During this period, 7.2 GB of data can be distributed (20 [Mbps] × (48 × 60) [sec] / 8).

  In the time zone “02:24 to 02:56”, data distribution according to the data distribution information “d05” can be performed with the remaining bandwidth (70 Mbps) as the assumed TP. Since 7.2 GB has been transmitted by time “02:24”, the remaining data can be transmitted in 21 minutes when the remaining 10.8 GB is distributed at 70 Mbps.

  In this way, a distribution schedule is created in which the assumed TP is set to 20 Mbps in the first half and the assumed TP is set to 70 Mbps in the second half. The content of the data distribution information “d05” in the distribution schedule information 134a is updated according to the determined distribution schedule.

  FIG. 50 is a diagram illustrating an example of the distribution schedule information after the distribution schedule of the data distribution information “d05” is determined. As shown in FIG. 50, when the transmission rate control is performed, if the remaining bandwidth of the NW link on the route is smaller than the bandwidth (80 Mbps) of the BL on the route of the data distribution information “d05”, transfer is possible with the remaining bandwidth. Data transfer is started at an assumed TP. When the other data transfer is completed and the remaining bandwidth of the NW link on the route increases, the data transfer with the assumed TP increased from that time. In the distribution schedule information 134a, there are a plurality of assumed TPs, required times, scheduled start times, and scheduled end times in the data distribution information "d05".

When the distribution schedule of the data distribution information “d05” is determined, the NW link use information is also updated.
FIG. 51 is a diagram illustrating an example of the NW link use information of the NW link “L4” after the distribution schedule of the data distribution information “d05” is determined. As shown in FIG. 51, for the NW link “L4”, the remaining band is “0” in the time zone “01:36 to 02:24”. That is, the resources of the NW link are used without waste.

  As described above, in the third embodiment, the distribution schedule is created on the assumption that the data transmission rate is dynamically changed, so that efficient data distribution using the resources of the NW link can be performed efficiently. Is

  As described above, the embodiment has been exemplified, but the configuration of each unit described in the embodiment can be replaced with another having the same function. Further, other arbitrary components and steps may be added. Further, any two or more configurations (features) of the above-described embodiments may be combined.

DESCRIPTION OF SYMBOLS 1 Data provision device 2a-2g Distribution destination device 3a-3g Relay device 4a-4c Group 10 Distribution schedule creation device 11 Storage unit 12 Operation unit 13 Distribution schedule

Claims (7)

On the computer,
A data providing device that holds data to be distributed, a plurality of destination devices to which the data is to be distributed, and a plurality of relay devices that relay communication between the data providing device and the plurality of destination devices The plurality of data providing devices, the plurality of distribution destination devices, and the communication speed of each of a plurality of links for communicatively connecting two devices among the plurality of relay devices based on topology information indicating the plurality of links. Among the destination devices, the destination devices having the same link with the lowest communication speed on the communication path with the data providing device are classified into the same group,
Among the destination devices belonging to the group, the first destination device having the highest communication speed of the directly connected link is specified,
Transmitting the data from the data providing device to the first destination device in the group, and then from the first destination device to a second destination device other than the first destination device in the group Creating a distribution schedule for the data to transmit the data;
A distribution schedule creation program that executes processing. The computer further comprises:
In the case where there are a plurality of the second destination devices belonging to the group, a second distribution in which a link having the lowest communication speed on a communication path with the first destination device among the second destination devices is common Classify the destination devices into the same subgroup,
The third destination device having the highest communication speed of the directly connected link is specified among the second destination devices belonging to the subgroup,
In creating the distribution schedule, the data is transmitted from the data providing device to a first distribution destination device in the group, and then the third distribution destination device in the subgroup is transmitted from the first distribution destination device. Transmitting the data from the third destination device to a fourth destination device other than the third destination device in the sub-group. Create a schedule,
The distribution schedule creation program according to claim 1. In the creation of the distribution schedule, when there are a plurality of groups, the lower the communication speed of the link having the lowest communication speed on the communication path between the distribution destination device and the data providing device belonging to each of the groups, the lower the communication speed is. Creating the distribution schedule so that transmission of the data to the first distribution destination device in the group is performed with priority.
The distribution schedule creation program according to claim 1. In the creation of the distribution schedule, when there are a plurality of data to be distributed, for each of the plurality of data, the total amount of data transmitted until the data is distributed to all distribution destination devices requesting data distribution. Calculating the total communication amount, and creating the distribution schedule so that transmission of the data is performed with higher priority for data having the larger total communication amount,
A distribution schedule creation program according to any one of claims 1 to 3. In the creation of the distribution schedule, for each of the distribution destination devices, a communication speed when receiving data distribution from the data providing device or another distribution destination device as a distribution source of the data is determined, and the communication speed is determined. In each of the links on the communication path between the data providing apparatus or another distribution destination apparatus serving as a data distribution source, the communication speed of data distribution using the link is subtracted from the communication speed of the link. Calculate the bandwidth and determine the communication speed according to the remaining bandwidth of the link with the least remaining bandwidth for each of a plurality of time zones,
A distribution schedule creation program according to any one of claims 1 to 4. Computer
A data providing device that holds data to be distributed, a plurality of destination devices to which the data is to be distributed, and a plurality of relay devices that relay communication between the data providing device and the plurality of destination devices The plurality of data providing devices, the plurality of distribution destination devices, and the communication speed of each of a plurality of links for communicatively connecting two devices among the plurality of relay devices based on topology information indicating the plurality of links. Of the destination devices, the destination devices having the same link with the lowest communication speed on the communication path with the data providing device are classified into the same group,
Among the destination devices belonging to the group, the first destination device having the highest communication speed of the directly connected link is specified,
Transmitting the data from the data providing device to the first destination device in the group, and then from the first destination device to a second destination device other than the first destination device in the group Creating a distribution schedule for the data to transmit the data;
Delivery schedule creation method. A data providing device that holds data to be distributed, a plurality of destination devices to which the data is to be distributed, and a plurality of relay devices that relay communication between the data providing device and the plurality of destination devices A storage unit for storing topology information indicating a relationship and a communication speed of each of a plurality of links for communicatively connecting two devices among the data providing device, the plurality of distribution destination devices, and the plurality of relay devices; ,
Based on the topology information, among the plurality of distribution destination devices, a distribution destination device having a link having a lowest communication speed on a communication path between the data providing device and the common distribution destination devices is classified into the same group, Among the distribution destination devices belonging to the group, the first distribution destination device having the highest communication speed of the directly connected link is specified, and the data is transmitted from the data providing device to the first distribution destination device in the group. An operation unit that creates a distribution schedule of the data so that the data is transmitted from the first distribution destination device to a second distribution destination device other than the first distribution destination device in the group.
A distribution schedule creation device having:

Images